Behavioral flexibility and the ability to respond appropriately to anthropogenic cues that signal potential threats or rewards may promote success of wild animals in the urban environment. Here, we examine the behavioral responses to anthropogenic cues of free-living American crows (Corvus brachyrhynchos), an opportunistic scavenger and common urban exploiter. Specifically, we tested the hypotheses that crows would more readily approach (1) novel objects and anthropogenic food when in close proximity to trash cans, and (2) paper bags with hidden food rewards when closer to a McDonald’s restaurant, due to their associations of food with these cues. In addition, we examined the preference of crows for bags marked with a McDonald’s logo versus unmarked bags, hypothesizing that crows closer to McDonald’s would be more likely to approach the labeled bag because of its familiarity. Consistent with our expectations, crows exhibited a lower approach latency to anthropogenic food and novel objects in close proximity to trash cans. Likewise, they were more likely to approach paper bags when in proximity to a McDonald’s restaurant, but they showed no preference for bags with a McDonald’s logo. Overall, we found evidence that the foraging behavior of crows varied in the presence of specific anthropogenic cues and that these responses depended on their probable cue familiarity. Their ability to use anthropogenic cues specific to the local environment, combined with their exploitation of human food resources, likely contributes to the success of the American crow in the city.
Data for Corvus brachrhynchos behavioral responses to food and anthropogenic objects/ cues. Experiments were conducted along an urban-rural gradient in Oneida County, NY, USA.
Contact Andrea Townsend ([aktownse@hamilton.edu] with any questions. This manuscript has been accepted for publication
Merz, M. R., S. Cote, R. Weinberg, T. Malley and A. K. Townsend. "Can I have fries with that? Context-dependent foraging behavior in urban and rural American crows." Behavioral Ecology, Volume 36, Issue 1, January/February 2025, arae098, https://doi.org/10.1093/beheco/arae098
Data files include:
Data files include:
File: experiment1_ImperviousSurface_TrashBin.csv
Variables
Treatment_Type: whether or not a trash can is present (Y/N)
Percent_impervious_50m: The percentage of impervious surface within a 50m radius around each test site
Number_individ_at_start: The number of crows present at the start of each trail
Bait_eaten: The number of pieces of food consumed during the trial
seconds_to_approach: The number of seconds that elapsed before a crow first contacted the bait in each trial
File: experiment2_McDonalds.Bag.csv
Variables
Distance.to.McDonalds(km): Distance from the experimental trial to the nearest McDonald's restaurant (in km)
impervious surface_50m: The percentage of impervious surface within a 50m radius around each test site
Bag.Preference (C= control; M = McDonalds): Bag first closely approached by crows in a given trial, where control bags were unmarked and McDonald's bags had a McDonald's logo. NA: Not Applicable; birds did not descend to the ground or approach either bag during the trial period.
descended to bait (Yes/No): Whether or not crows descended to the ground during the trial period
Experiment 1: Foraging behavior, urbanization, and presence of trash cans
Behavioral responses to novel food and objects were recorded along an urbanization gradient at sites in Utica, Kirkland, New Hartford, and Rome, New York (USA) between October 2014 and March 2015 and September 2020 and April 2021 (Fig. 1). Sites were selected in parking lots in urban centers and on roadsides along agricultural fields in rural areas. Agricultural roadsides had minimal traffic; no cars drove past for the entire duration of most roadside trials (e.g., Supplementary materials, Video S1). To characterize the degree of urbanization, we created buffers with a 50-meter radius surrounding each data collection site in QGIS 3.16. We calculated the percent impervious surface cover based on data from the 2016 National Land Cover Database (Percent Urban Imperviousness dataset), using the Zonal Statistics tool in QGIS to quantify the percentage of pixels covered by impervious surfaces within each buffer. We chose a 50-meter radius because it is a spatial scale relevant to the microhabitat and local associations of individuals in the immediate vicinity of each testing location. The percentage of impervious surface cover in these trials ranged from 0 – 87% (mean: 31.96 ± 4.99%; Supplementary materials, Fig. S1a). Qualitatively, sampling points in and surrounding Utica, NY (population = 64K; population density 1500 people/km2), were relatively urban; points in the village of Clinton, NY (population = 1.7K; population density 1000 people/km2), were intermediate, and points surrounding Kirkland, NY (population = 10K; population density 116 people/km2), were relatively rural (Fig. 1).
We chose sites for our trials along pre-selected routes, with exact locations based on the presence of unique groups of crows along these routes. At the start of all trials, the nearest crow was 15-25m from the experimental setup, a distance that we confirmed with a rangefinder (Nikon Monarch Gold). At this distance, birds did not fly away when we set up the experiment but were still close enough to observe the experimental setup. The experimental setup is shown in Fig. 2. We defined two contexts in which to compare the foraging behavior of crows encountered along these routes: sites at which a novel object and anthropogenic food items were placed within two meters of a trash can or dumpster ("trash treatment”) and sites at which a novel object and food items were placed out of view of the nearest trash can or dumpsters (i.e., at least 75m from the nearest trash can or dumpster; “control treatment”). At each site, we stepped out of the car, placed a video camera (the GoPro Hero 4 Black) on the pavement, and started recording with the GoPro. The GoPro served as the novel object in our experiment while simultaneously recording the crows’ foraging behavior. We tossed 20 cheese puffs, a brightly colored food reward that is highly attractive to crows (Caffrey 2002) but with which they are likely to be unfamiliar, in the air in front of the GoPro to get the attention of the crows and returned to our car, moving it at least 20 meters away from the experimental setup to observe and video record the responses of crows (Supplementary materials, Videos S1 and S2). Latency to approach (defined as the number of seconds that elapsed before any crow in the group made first contact with a food item, with the timing beginning immediately after parking our car >20m from the experimental setup), the number of crows present at the start of the trial, amount of bait consumed, and GPS location were recorded for each trial. Trials were recorded for 20 minutes or until all the bait was gone. If no crows approached the bait within 20 minutes but were present for the entire 20 minutes, we stopped the experiment and recorded 20 minutes as the latency to approach the bait for that trial. If crows dispersed from the area completely without having ever approached the experimental setup, the trial was discontinued and data were not recorded.* *Likewise, trials with human disturbance (e.g., pedestrians or vehicles; *n *= 6 trials) were discontinued and excluded from the sample. All trials were conducted within 2.5 hours of sunrise.
In total, ten sampling sites were established on roads along agricultural fields, seven on Hamilton College campus, and 20 in urban centers (Fig. 1). To minimize the likelihood that crows participated in multiple trials, agricultural sites were established ≥ 0.4 km apart and urban sites were located in separate parking lots at least 100m apart. Campus sites were established on known crow territories, and we only performed one test (i.e., trash treatment or control) within each territory.
Experiment 2: Urbanization and bag preference (labeled vs. unlabeled)
To evaluate the extent to which crows use the McDonald’s logo as a cue for food resources (i.e., their preference for labeled bags vs. unmarked bags), we collected data in Kirkland, Rome, Oneida, Utica, and New Hartford, New York from October 2019 to March 2020 and in November 2023. Twenty-four sites were established at pre-selected locations, at specific distances (0.5 km, 1.0 km, and 1.5 km) from six different McDonald’s restaurants. Sites were pre-selected using Google Earth in parking lots of non-food establishments (e.g., automotive garages, gas stations, health centers, and schools) or street sides in residential neighborhoods. Seventeen additional locations were randomly selected in parking lots in urban centers or along roadsides in rural areas, based on the presence of unique groups of crows for testing. Crows were present at only 18 of the 24 pre-selected sites; therefore, the final sample included only 18 preselected sites and 17 randomly selected sites (*n *= 35 sites). Only one trial was conducted at each location to minimize the likelihood that crows participated in multiple trials. As above, degree of urbanization was characterized as the percent impervious surface cover within a 50m radius around each testing site. The percent impervious surface cover in these trials ranged from 0.13 – 85.13% (mean: 44.65 ± 4.42%; Supplementary materials, Fig. S1b)
Before each trial, one McDonald’s cheeseburger was placed inside a plain brown paper bag and one inside a McDonald’s bag. The burgers were immediately frozen after purchase and thawed overnight in the refrigerator prior to each trial. Bags containing burgers (one marked, the other unmarked) were placed on the ground two meters apart from one another, within five meters of a perched crow. Bags were oriented so that birds could not see their contents without approaching them closely (Fig. 3; Supplementary materials, Video S3). Observers then returned to their car, moving it at least 20 meters away from the experimental setup to observe the trial. The order that the bags were placed on the ground was randomly selected using an online sequence generator (random.org/lists). The bag (marked or unmarked) first approached by a crow was recorded at each location. Birds were allowed to remove and eat the burger from the first bag that they approached; trials were interrupted and bags collected before they were able to approach the second bag. Trials were ended if the crows did not descend to either bag within twenty minutes.